transformers-js

Use Transformers.js to run state-of-the-art machine learning models directly in JavaScript/TypeScript. Supports NLP (text classification, translation, summarization), computer vision (image classification, object detection), audio (speech recognition, audio classification), and multimodal tasks. Works in Node.js and browsers (with WebGPU/WASM) using pre-trained models from Hugging Face Hub.

Quality

77%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./plugins/hugging-face/skills/transformers.js/SKILL.md

Quality

Content

72%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-structured skill with executable examples for every task and a clean one-level reference layout, weakened mainly by content redundancy that inflates the body and by missing validation guidance for batch operations. Pruning duplicate tables/lists and adding a verify step for batch use would lift the weaker dimensions.

Suggestions

Remove the duplicated Hugging Face Hub URL material: keep it in one place (the "Finding and Choosing Models" table) and have "Model Selection" link there instead of repeating the links.

Drop or merge the closing "Quick Reference: Task IDs" table with the "Supported Tasks" section to eliminate the task-list duplication.

Add an explicit validation/verification step to the Batch Processing example (e.g., checking that the result array length matches the input array length, or handling per-item errors) so the batch workflow has a feedback checkpoint.

Dimension	Reasoning	Score
Conciseness	The ~635-line body is mostly efficient executable code, but it repeats material: the Hugging Face Hub URL list appears both in "Model Selection" and again as a full table in "Finding and Choosing Models", the task catalogue is duplicated between "Supported Tasks" and the closing "Quick Reference: Task IDs" table, and "Best Practices" overlaps "Performance Tips" and "Memory Management". It is tighter than concept-explaining fluff (not a 1) but carries clear redundancy that could be pruned (not a 3).	2 / 3
Actionability	Nearly every section gives complete, copy-paste-ready executable JavaScript — installation, pipeline creation per task, env configuration, progress callbacks, and error handling — with concrete model IDs and option objects, matching the fully-executable anchor; minor variable-name slips (e.g., `classifier.dispose()` where `pipe` was created) do not undermine overall copy-paste readiness.	3 / 3
Workflow Clarity	The core install→create-pipeline→run→dispose pattern is clear and the model-selection example is a numbered sequence, but the "Batch Processing" section shows calling the classifier on an array with no validation or verification step; per the guideline, missing validation in batch operations caps workflow clarity at 2 rather than 3.	2 / 3
Progressive Disclosure	The body is an overview with well-signaled, one-level-deep references — "→ [Configuration Reference](./references/CONFIGURATION.md)", CACHE.md, PIPELINE_OPTIONS.md, TEXT_GENERATION.md, MODEL_ARCHITECTURES.md, EXAMPLES.md — all of which exist as real files, plus a dedicated "Reference Documentation" section listing them, matching the clear-overview-with-one-level-deep-references anchor.	3 / 3
	Total	10 / 12 Passed

Description

82%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that names concrete capabilities and natural trigger terms across a clearly distinct niche, but it omits an explicit "Use when..." trigger clause, capping completeness. Adding a one-sentence trigger clause would round it out.

Suggestions

Append a "Use when..." clause naming the situations that should activate this skill (e.g., running ML models client-side in JS, in-browser inference, or when the user mentions Transformers.js/Hugging Face models in JavaScript).

Consider adding a couple of layperson trigger terms (e.g., "in-browser AI", "on-device inference") alongside the technical ones to broaden natural activation.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capability areas across modalities — "text classification, translation, summarization", "image classification, object detection", "speech recognition, audio classification", and "multimodal tasks" — matching the anchor for listing several specific concrete actions rather than naming only a domain.	3 / 3
Completeness	It clearly answers "what" (run ML models in JS across NLP/CV/audio/multimodal, in Node.js and browsers) but contains no "Use when..." clause or equivalent explicit trigger guidance, so per the judging guideline completeness is capped at 2.	2 / 3
Trigger Term Quality	Covers natural terms users would actually say — "machine learning models", "JavaScript/TypeScript", "Node.js", "browsers", "speech recognition", "image classification", "Hugging Face Hub" — giving good coverage of likely trigger phrasing rather than only technical jargon.	3 / 3
Distinctiveness Conflict Risk	"Use Transformers.js to run state-of-the-art machine learning models directly in JavaScript/TypeScript" defines a clear, specific niche that is unlikely to trigger for unrelated skills, matching the distinct-niche anchor; not a 2 because it is far more specific than a generic "works with document files".	3 / 3
	Total	11 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (636 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: openai/plugins
Path: plugins/hugging-face/skills/transformers.js/SKILL.md
Commit: 11c74d6

Reviewed: 17 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.